Transactions of Japan Society of Kansei Engineering
Online ISSN : 1884-5258
ISSN-L : 1884-0833
NOUN BASED KANSEI IMAGE RETRIEVAL SYSTEM BY THE USE OF MTS METHOD
Shuichi AKIYAMAKensuke SUZUKIAndreas GEHRMANNYoshimitsu NAGAISyohei ISHIZU
Author information
JOURNALS FREE ACCESS

2009 Volume 8 Issue 4 Pages 1171-1178

Details
Abstract

Image retrieval systems are widespread in areas such as online shopping, etc. Usually images are retrieved according to the annotated texts. This text image retrieval needs a lot of load of time and effort to annotate the images. As an effective retrieval system, Kansei image retrieval system is proposed. By Kansei retrieval system we can retrieve images based only on adjective but not on noun. This fact leads limited use of Kansei retrieval systems in online shopping area. One of our main aims of this paper is to propose a noun and adjective based image retrieval system. In order to propose noun based image retrieval method, we apply MTS method, since MTS method is a useful pattern information processing method by the use of Mahalanobis distance. We construct image retrieval system for apparel products, and investigate accuracy of noun retrieval and fitness of noun and adjective based retrieved image.

Information related to the author
© Japan Society of Kansei Engineering
Previous article Next article
feedback
Top